Reducing strong noise with the same frequency in magnetostrictive guided waves using improved orthogonal matching pursuit

Noise often masks defect-related signals during the assessment of bridge cables based on guided waves (GWs) to reduce the accuracy of detection. Traditional methods of denoising struggle to provide satisfactory results when using GWs under interference by highly dense noise of the same frequency. In this context, this study optimises the traditional orthogonal matching pursuit algorithm by using a novel condition for terminating its iterations and improving its efficiency of search for atoms, and uses it to denoise GWs under interference by high-density noise with the same frequency to detect broken wires in bridge cables. The results showed that highly dense noise could be reduced to achieve accurate results when the threshold of the correlation coefficient between the best-matching atoms of the reference signal and the target signal within a range of 0.5 to 0.7 was used as the condition for terminating iterations. The proposed condition for terminating iterations is more adaptable to signals with different signal-to-noise ratios (SNRs) than conditions based on the number of iterations, residual energy ratio, and adjacent residual ratio. The proposed method also exhibited a superior capability of noise reduction in comparison with the WT method of denoising and the VMD-linked wavelet method.